4471
`
`Structure of Bis(methylguanidinium) Monohydrogen
`Orthophosphate. A Model for the Arginine-Phosphate
`the Active Site of Staphylococcal
`Interactions at
`Nuclease and Other Phosphohydrolytic Enzymes1
`F. A. Cotton,*2a V. W. Day, E. E. Hazen, Jr.,*2a S. Larsen,2b and S. T. K. Wong
`Contribution from the Departments of Chemistry, Texas A&M University,
`77843, and University of Nebraska,
`College Station, Texas
`68508. Received February 6, 1974
`Lincoln, Nebraska
`
`Abstract: A compound has been found which provides an excellent model of certain essential features of the ar-
`ginyl-phosphate interactions in the complex consisting of Staphylococcus nuclease, deoxythymidine 3',5'-diphos-
`phate, and calcium ion. The crystal structure of this substance, bis(methylguanidinium) monohydrogen ortho-
`space group, Fddl-C2lls; unit cell
`phosphate, [CNsNHCCNtLhkPO3OH, has been determined. Crystal data:
`16 for the formula unit (C2N3H5)2HP04; dcaiCd =
`dimensions a = 23.608 (3), b = 24.113 (5), c = 7.917 (1) Á; Z =
`1.430 g cm-3. Using Zr-filtered Mo Ka radiation a total of 5428 reflections having  -1 sin   <1.030
`1.436, dob,d =
`were measured with an automated diffractometer. Using 4465 reflections adjudged to the statistically significant,
`the structure was solved by Patterson and Fourier methods and refined by full-matrix least squares to final unit-
`weighted and weighted residuals of 0.049 and 0.048, respectively. The crystallographically independent guanidyl
`groups are planar and each forms two hydrogen bonds to the HPQ42- ion through separate N-H groups. One
`phosphate oxygen atom participates in two hydrogen bonds, one from each guanidyl. The overall arrangement is
`very similar to, though not precisely the same as, that in the enzyme-inhibitor complex and provides an excellent
`model for the latter. The HPCV- ions form hydrogen bonded pairs related by a twofold axis. The O-H- -O dis-
`tances here are relatively short, 2.544 and 2.503 Á, but it would appear that the hydrogens must be considered to be
`disordered rather than symmetrically located. The P-0 distances of 1.514, 1.524, 1.556, and 1.567 A are values
`that might be more typically expected for an H2P04- ion and may be a reflection of the hydrogen-bonding effect of
`the guanidyl ions.
`
`Only in recent years has it become apparent that the
`guanidyl groups of arginine residues play an im-
`portant role in binding and possibly even more active
`roles at the functional sites of both enzymes and non-
`catalytic proteins. Limiting reference only to those
`cases where it has been reasonably demonstrated that
`the chemical modification of arginine does occur at the
`binding or functional site of the protein molecule, func-
`tionally active arginine residues have been found in E.
`coli alkaline phosphatase,3 yeast
`inorganic pyrophos-
`phatase,4 5lactate dehydrogenase,6 D-amino acid oxi-
`dase,6 pepsin,7 ribonuclease Ti,8 carboxypeptidases A9
`and B,10 and antibody combining sites directed against
`haptens containing such anions as arsonate, phos-
`phonate, and carboxylates.11-16 Of particular interest
`
`(1) This study was supported by Grant GM13300 from the National
`Institute of General Medical Sciences, National
`Institutes of Health.
`The structure has been briefly described in a preliminary note:
`F. A.
`Cotton, E. E. Hazen, Jr., V. W. Day, S. Larsen, J. G. Norman, Jr.,
`S. T. K. Wong, and K. H. Johnson, J. Amer. Chem. Soc., 95, 2367
`(1973).
`(2) Address correspondence to these authors at
`the Department of
`Chemistry, Texas A&M University, College Station, Texas
`77843.
`(b) On leave, 1970-1971, from the Technical University of Copenhagen.
`(3) F. J. M. Daemen, Fed. Proc., Fed. Amer. Soc. Exp. Biol., 32, 473
`(1973).
`(4) B. S. Cooperman and N. Y. Chiu, Biochemistry, 12, 1676 (1973).
`(5) P. C. Yang and G. W. Schwert, Biochemistry, 11, 2218 (1972).
`(6) A. Kotaki, M. Harada and K. Yagi, J. Biochem. (Tokyo), 60,
`592(1966).
`(7) W.-Y. Huang and J. Tang, J. Biol. Chem., 247,2704 (1972).
`(8) A. Takahashi, J. Biochem. (Tokyo), 68, 659 (1970).
`(9) J. F. Riordan, Biochemistry, 12, 3915 (1973).
` . M. Weber and M. Sokolovsky, Biochem. Biophys. Res.
`(10)
`Commun., 48, 384 (1972).
`(11) A. L. Grossberg and D. Pressman, Biochemistry, 7,272 (1968).
`(12)  . H. Freedman, A. L. Grossberg, and D. Pressman, Bio-
`chemistry, 7,1941 (1968).
`
`are those cases where the chemical modifications in solu-
`tion can be correlated with the results from cr>: al
`structure analyses. Thus, the single arginine which is
`shown by chemical modification to be at the active site
`in carboxypeptidase A9 is very probably Arg-145 which
`Lipscomb, et a/.,16 in their crystallographic studies,
`have found to bind the terminal carboxylate of peptides.
`Similarly, the estimation of three essential arginines per
`subunit of lactate dehydrogenase by Yang and Schwert
`from their chemical modification studies6 correlates
`very neatly with the recent observations from Ross-
`mann’s and Kaplan’s laboratories.
`These observations
`indicate arginines-101, -109, and -171 are located at the
`active site in the crystal structure of the abortive lactate
`dinucleotide-
`adenine
`dehydrogenase-nicotinamide
`pyruvate ternary complex. The first of these arginines
`bridges the pyrophosphate linkage in the coenzyme,
`and the other two interact with the substrate.17
`Our direct determination of the high resolution crystal
`structure of the ternary complex of the Staphylococcal
`nuclease with its potent competitive inhibitor,
`thymi-
`dine 3',5'-diphosphate and calcium ion has revealed
`that the 5'-phosphate of the inhibitor forms two hydro-
`gen bonds each to the guanidinium ions of arginines-
`(13)  . H. Freedman, A. L. Grossberg, and D. Pressman, J. Biol.
`Chem., 243,6186(1968).
`(14) G. L. Mayers, A. L. Grossberg, and D. Pressman,
`Immuno-
`chemistry, 9,196 (1972).
`(15) G. L. Mayers, A. L. Grossberg, and D. Pressman, Immunochem-
`istry, 10,37(1973).
`(16) W. N. Lipscomb, G. N. Reeke, Jr., J. A. Hartsuck, F. A. Quio-
`cho, and P. H. Bethge, Proc. Roy. Soc., Ser. B, 257,177 (1970).
`(17) M.
`J. Adams, M. Buehner, K. Chandrasekhar, G. C. Ford,
`M. L. Hackert, A. Liljas, M. G. Rossmann, I. E. Smiley, W. S. Allison,
`J. Everse, N. O. Kaplan, and S. S. Taylor, Proc. Nat. Acad. Sci. U. S.,
`70,1968(1973).
`
`Cotton, et al.
`
`/ Bis(methylguanidinium) Monohydrate Orthophosphate
`
`Downloaded via PENNSYLVANIA STATE UNIV on August 2, 2020 at 20:31:57 (UTC).
`
`See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
`
`Merck Exhibit 2208, Page 1
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`4472
`
`CH3
`
` - —N
` .
`>P
`0* X0— H—N''"
`IH
`
`CNH,
`
`1
`
`NH2
`
`nh2
`I/Cx
`
`HN
`
`NCH,—
`
`I
`
`I
`
` 
`
`I
`
`I
`
`H
`/¿— —   h
`Ó.
`^P'
`,CN
` * X0— —N/
`I
`H2C—
`
`I
`
`-CX
`
`!
`
`h3cn
`
`H (
`
`NH
`IH
`,0-H-Nx
`X
`JPX
`/CNHl
`0— —N
`CT
`
`CH3
`
`I
`
`IH
`
`The search for models was begun with substances ob-
`tainable from solutions containing methylguanidinium
`ion [CH3NHC(NH2)2]+ (MGD) which closely resembles
`the end of the side chain of an arginyl residue and phos-
`phate ions, HkP04”-3. The first substance obtained
`(MGD)H2P04, which exhibits the
`and examined25 was
`type of double hydrogen-bonded interaction that is of
`interest, 1, but does not mimic the entire arrangement
`seen in the enzyme-inhibitor complex, 2.
`When crystals of a second methylguanidinium phos-
`phate, (MGD)2HP04, were obtained, it was considered
`worthwhile to carry out another structure determination
`with the objective of finding a more complete facsimile
`of
`the enzyme-substrate arrangement, 2. This ob-
`jective has been accomplished, since, as shown here,
`(MGD)2HP04 contains the arrangement, 3, which very
`closely resembles 2, though it does not precisely dupli-
`cate it.
`
`Experimental Section
`An aqueous solution containing equimolar quantities of methyl-
`guanidinium sulfate (Eastman Organic Chemicals) and Ba(OH)2
`then filtered
`was stirred overnight to precipitate BaS04 which was
`off. Half an equivalent amount of phosphoric acid was added to
`the filtrate to form an aqueous solution of bis(methylguanidinium)
`monohydrogen phosphate (pH 7.5). After evaporating this solution
`to near dryness, ethanol was added until the solution became turbid.
`Large single crystals of bis(methylguanidinium) monohydrogen
`phosphate grew over a period of several days.
`Anal. Caled for
`19.67; N,
`[(NH2)2CNHCH3MHP04): C,
`34.42;  , 7.02. Found: C, 19.42; N, 34.19,  , 7.03.
`A spherical specimen 0.70 mm in diameter was ground from a
`larger crystal and glued to the end of a glass fiber with a tip diam-
`eter of 0.10 mm.
`Precession photographs, used to determine a preliminary set of
`indicated orthorhombic, mmm, symmetry. The
`lattice constants,
`systematically absent reflections were
`those uniquely required by
`the noncentrosymmetric space group, Fddi-Ctvn This choice was
`fully supported by the positive results of sensitive tests for piezo-
`electricity and by the subsequent structure determination. The
`full-circle goniom-
`crystal was accurately centered on a Syntex PI
`eter and a total of 15 reflections, chosen to give a good sampling
`of reciprocal space and diffractometer settings (IduoKa > 50°),
`were used to align the crystal and calculate angular settings for each
`reflection. A least-squares refinement of the diffraction geometry
`for these 15 reflections, recorded at the ambient laboratory temper-
`radiation (X(Mo   ) 0.71069 Á) gave
`ature of 21 ± 1
`° with Mo   
`24.113 ± 0.005, and
`the lattice constants a =
`23.608 ± 0.003, b =
`7.917 ± 0.001 A. A unit cell content of 16 bis(methylgua-
`c
`nidinium) monohydrogen phosphate molecules gives a calculated
`density of 1.436 g/cm3, in good agreement with the observed den-
`sity of 1.430 g/cm3, measured by flotation in a mixture of dichloro-
`methane and carbon tetrachloride.
`Intensity measurements utilized Zr-filtered Mo   
`radiation and
`the  -2  scanning technique with a 3° takeoff angle and a standard-
`focus X-ray tube on a computer controlled Syntex PI diffractom-
`eter. A scanning rate of 3°/min was employed for the scan be-
`tween 2  settings 1.0° above and below the calculated   
`doublet
`0.70926 and (  2) 0.71354 A) of each reflection ex-
`values ( (   )
`cept for those reflections having 83.3° < 2  < 94.1 ° where a 2°/min
`lasting half
`scanning rate was used. Background counts (each one
`the total scan time) were taken at both ends of the scan range. A
`total of 5428 independent reflections having (sin 9/ ) < 1.030 (four
`times the number of data in the limiting Cu   
`sphere) were mea-
`sured in concentric shells of increasing 29 containing approximately
`1400 reflections each. The six standard reflections, measured
`every 300 reflections as a monitor for possible disalignment and/or
`deterioration of the crystal, gave no indication of either.
`The linear absorption coefficient of the crystal for Mo   
`radia-
`tion is 0.26 cm"1, yielding µ  of 0.09 for the spherical crystal used.
`Since the absorption of X-rays by a spherical crystal having µ 
`0.09 is essentially independent of scattering angle, no absorption
`correction was made, and the intensities were reduced to relative
`squared amplitudes, |F0|2, by means of standard Lorentz and polar-
`ization corrections.
`
`=
`
`=
`
`35 and -87.18- 20 Additional hydrogen bonds occur
`between these guanidinium ions and other parts of the
`enzyme molecule to effectively lock the 5'-phosphate
`rigidly in the active site. Thus, our structural observa-
`tions suggest a rather unique and highly specific func-
`tional role for these two arginine residues, a notion that
`is strongly reinforced by the observation of Chaiken
`and Anfinsen that the replacement of Arg-35 by either
`lysine or citrulline in semisynthetic variants of
`the
`nuclease completely abolishes enzymatic activity.21,22
`Since the chemical modification studies have indicated
`the presence of arginines at active sites of other phopho-
`hydrolytic enzymes, namely, alkaline phosphatase,3
`inorganic pyrophosphatase,4 and ribonuclease Ti,18
`there may well be a general subclass of enzymes in-
`volved in phosphate metabolism having arginines at
`their active sites.
`These structural and chemical observations as to the
`importance of
`specificity and functional
`the guan-
`idinium-phosphate interactions in the Staphylococcal
`nuclease lead to a search for simpler model systems.lb 23
`There are two purposes in examining model systems:
`(1) to confirm the plausibility of the overall interpreta-
`tion of the enzyme-inhibitor interaction obtained from
`fitting to the electron density maps of
`model
`the
`nuclease-deoxythymidine 3 ',5 '-diphosphate-Ca2+ com-
`plex, and (2) to obtain accurate structure parameters
`refined understanding of the
`which can provide a more
`interactions than that obtainable from the enzyme struc-
`ture itself.
`(18) Historically protein chemists have referred to such interactions
`“salt
`such groups bearing formal charges as
`between two
`linkages”
`“salt bridges.” Since this terminology was
`first introduced,
`there
`or
`has been ample evidence from the crystal structures of smaller systems
`that interactions such as these normally occur
`via hydrogen bonds and
`since extensive charge delocalization is characteristic of most charged
`the use of this rather artificial
`groups found in proteins,
`terminology
`when the hydrogen bonding groups both have formal charges, could
`well be discontinued.
`(19) A. Amone, C. J. Bier, F. A. Cotton, V. W. Day, E. E. Hazen,
`Jr., D. C. Richardson, J. S. Richardson, and A. Yonath, J. Biol. Chem.,
`246,2302(1971).
`(20) F. A. Cotton, C. J. Bier, V. W. Day, E. E. Hazen, Jr., and S.
`Larsen, Cold Spring Harbor Symp. Quant. Biol., 36,243 (1971).
`(21) I. M. Chaiken and C. B. Anfinsen, J. Biol. Chem., 246, 2285
`(1971).
`(22) In preliminary results from this laboratory, phenyl glyoxal has
`been found to modify the arginine residues of the nuclease with the
`loss of enzymatic activity.
`concurrent
`(23) F. A. Cotton, V. W. Day, E. E. Hazen, Jr., and S. Larsen, J.
`Amer. Chem. Soc., 95,4834 (1973).
`
`Journal of the American Chemical Society / 96:14 / July 10, 1974
`
`Merck Exhibit 2208, Page 2
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`Of the 5428 reflections examined, 963 were rejected as objectively
`unobserved by applying the rejection criterion, I <  (/), where  (/)
`is the counting statistics standard deviation in the observed inten-
`sity computed from
`
`=
`
`=
`
`—
`
`-
`
`-
`
`=
`
`=
`
` (/) =
`(Ct + /e25y-
`Ct being the total count from scanning, k the ratio of scanning time
`1), and B the total
`to total background time (in this case k =
`The remaining 4465 observed intensities were
`background count.
`used in the determination and refinement of the structure.
`Structure determination was achieved through a combination of
`the heavy-atom technique, difference Fourier syntheses, and least-
`squares refinement. The wholly straightforward interpretation of
`the Patterson synthesis of the 697 |F0|2 data having (sin 0/X) < 0.52
`placed the phosphorus atoms in 16-fold general positions (0,0,0;
`x,lU + y'U + z;
`o,1/2, V2; 7í,o,Ví; Vid/a-p) + (x,y,z; s,y¿; 74 -
`+  -,V 4
`— y,lU + z.
`These atomic coordinates, an isotropic
`thermal parameter, and a scale factor were varied in two cycles of iso-
`tropic full-matrix least-squares refinement.24 This resulted in a con-
`ventional unweighted residual, Ri = 0.499, for these low angle data.
`R!
`|F,||/2|F0|
`= 2||F0|
`A difference electron density map at this stage revealed the loca-
`tions of the four phosphate oxygen atoms. Two cycles of least-
`squares refinement varying the scale factor, atomic coordinates, and
`isotopic temperature factors for phosphorus and oxygen atoms,
`0.372. A second electron density differ-
`respectively, gave Ri
`ence map clearly revealed the remaining ten non-hydrogen atoms of
`the asymmetric unit, all of which lie in general positions.
`Iso-
`tropic full-matrix refinement using unit weighting for the 15 non-
`0.072, for 697 reflections. All of the
`hydrogen atoms gave R\
`then included in a fully anisotropic least-
`4465 reflections were
`k|Fc|)2 to give, with
`squares minimization of the function 2w(|F0|
`unit weighting (i.e., all w =
`0.063. This and all subse-
`1) Ri
`quent refinement cycles employed an anomalous dispersion cor-
`rection25 to the scattering factor of the phosphorus atom and a
`least-squares refinable extinction coefficient26 of the form F(x) =
`1/(1 + 2x)1/’-, where x = gland g refined to a final value of 0.36 X
`10~7. A Fourier difference synthesis based on the refined param-
`eters afforded direct evidence for the placement of all hydrogen
`atoms. Further unit-weighted full-matrix least-squares
`cycles
`used to refine hydrogen atoms isotropically and all other
`were
`0.049 and a conventional
`atoms anisotropically to give Ri
`weighted residual, Ri = 0.044.
`|2>(iF0|
`|Fc|)2/2w|F0S2)Vi
`R2 =
`l/ 2) were then calculated from
`Empirical weights (w =
`3
`= 2>„|F0|n =
`0
`
` 
`
`1.54 -
`
`0.20 X 10“ lF +
`0.23 X 10~3jF2 -
`0.35 X 10~6F3
`the an being coefficients derived from the least-squares fitting of the
`curve
`
`4473
`with standard deviations by Busing, Martin, and Levy; ortep-h,
`thermal ellipsoid plotting program by C. K. Johnson; mplane,
`plane calculation program from L. Dahl’s
`least-squares mean
`group.
`
`Results
`The final coordinates and anisotropic thermal param-
`eters for all atoms except hydrogen atoms are listed in
`Tables I and II, respectively;
`the refined positions and
`
`Table I. Atomic Coordinates in Crystalline
`Bis(methylguanidinium) Monohydrogen Phosphate"
`Atom6
`type
`
`105x
`
`Nx
`N2
`n3
`c,
`c2
`
`N,
`N2
`n3
`Cl
`c2
`
`9959 (9)
`9264(11)
`6717 (9)
`12657 (16)
`8655 (8)
`
`23115 (7)
`28807 (7)
`20656 (7)
`27274 (12)
`24204 (7)
`
`105y
`Cation I
`10244 (7)
`10208 (8)
`2542 (6)
`15602(11)
`7711 (6)
`Cation II
`1538 (7)
`8798 (6)
`10454 (7)
`-2734 (9)
`6942 (7)
`Anion
`-164(1)
`0
`P
`8675(1)
`-1514 (2)
`-2160(5)
`Ol
`4917 (5)
`-1066(7)
`1645 (2)
`5189 (6)
`o2
`-3632 (4)
`73(2)
`14016(4)
`o3
`-199 (2)
`o4
`5994 (4)
`9965 (5)
`" Figures in parentheses are the estimated standard deviations.
`Coordinate listed without standard deviation is symmetry re-
`1 Atoms numbered to agree with Figures 1-5.
`quired.
`
`104z
`
`-3507 (2)
`-6414(2)
`-4890 (3)
`-3417 (3)
`-4935 (3)
`
`1680 (3)
`2535 (3)
`1008 (3)
`2017 (5)
`1758 (2)
`
`isotropic thermal parameters of the hydrogen atoms are
`listed in Table III.27 The rule used in the atom num-
`bering scheme for bis(methylguanidinium) monohydro-
`gen phosphate is as follows. Atoms of the methyl-
`guanidinium ions are grouped according to cation. A
`numerical subscript is used to differentiate atoms of the
`same non-hydrogen element. For each hydrogen atom
`the subscript letter and first subscript number indicate
`the atom to which it is covalently bonded, while the
`second numerical subscript distinguishes among hydro-
`gen atoms attached to the same atom.
`A projection of one asymmetric unit is presented in
`Figure 1; each atom is numbered in conformity with
`Tables I-IX and each non-hydrogen atom is represented
`by an ellipsoid having shape, orientation, and relative
`size consistent with the thermal parameters listed in
`Tables II. Bond lengths and angles in the molecular
`skeleton are presented in Figure 2 and are listed along
`with their estimated standard deviations in Tables IV
`the dimensions of various interionic hydrogen
`and V;
`bonds are listed in Table VI. The equations of the mean
`planes that partially characterize important subgroup-
`ings of atoms within the asymmetric unit specified by
`the coordinates of Tables I and III are given in Table
`VII,27 and the displacements from these planes of the
`atoms constituting the asymmetric unit are
`listed in
`Tables VIII and IX.27
`
`(27) See paragraph at end of paper
`terial.
`
`regarding supplementary ma-
`
`n
`
`IIF I
`II* o|
`
`—
`
`| F ||
`I* c| [
`
`—
`
`3
`-- Vfl
`\F \
`¿^/un\L o|
`0
`The Fc values were calculated from the fully refined model using
`unit weighting. The final cycles of least-squares refinement uti-
`lized these weights and anomalous dispersion corrections for the
`phosphorus atom to refine hydrogen atoms isotropically and all
`other atoms anisotropically together with the scale factor and ex-
`tinction coefficient to give final values of 0.049 and 0.048 for Ri and
`Ri, respectively. During this last cycle of refinement no param-
`eter for non-hydrogen atoms shifted by more
`than 0.2 , with the
`average shift being 0.03 .
`The following computer programs were employed in this work:
`and sctft2, data reduction programs written by V. Day;
`magtap
`fordap, Fourier and Patterson synthesis program, a modified ver-
`sion of A. Zalkin’s original program; orflse,
`full-matrix least-
`squares refinement program, a highly modified version of Busing,
`Martin, and Levy’s original orfls; orffe, bond lengths and angles
`
`taken from D. T. Cromer and J. B.
`(24) Scattering factors were
`Mann, Acta Crystallogr., Sect. A, 24, 321 (1968).
`(25) D. T. Cromer, Acta Crystallogr., 18,17 (1965).
`(26) W. H. Zachariasen, Acta Crystallogr., 23, 558 (1967).
`
`Cotton, et al.
`
`j Bis(methylguanidinium) Monohydrate Orthophosphate
`
`Merck Exhibit 2208, Page 3
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`4474
`Table II. Anisotropic Thermal Parameters in Crystalline Bis(methylguanidinium) Monohydrogen Phosphate"
`Atom6
`type
`
`Bn
`
`Bzs
`
`Bn
`
`Bn
`
`Bn
`
`fl,= A2
`
`flu
`
`N,
`N,
`n3
`Cl
`c.
`
`Ni
`n2
`n3
`c,
`c2
`
`4.31(8)
`5.83 (11)
`5.17(9)
`6.81 (16)
`2.93 (5)
`
`2.31 (5)
`2.74 (5)
`2.50 (5)
`4.01 (9)
`1.98 (5)
`
`2.05 (5)
`2.22 (6)
`2.07 (4)
`2.97 (8)
`1.80 (4)
`
`1.99 (4)
`2.20(4)
`2.10(5)
`2.19 (6)
`2.01 (4)
`
`1.64(5)
`1.73 (5)
`2.09 (5)
`2.55 (8)
`1.73 (4)
`
`4.17 (8)
`3.24 (6)
`3.83 (8)
`5.62(14)
`2.33 (5)
`
`0.04 (5)
`-0.01 (6)
`-0.10(7)
`-0.08 (9)
`0.11 (5)
`
`-1.16(6)
`-1.16(6)
`-0.87 (5)
`-2.12(10)
`-0.19 (4)
`
`0.13 (4)
`0.26 (4)
`0.16(5)
`0.13 (7)
`0.27 (5)
`
`-0.01 (5)
`0.19(5)
`-0.38 (5)
`-0.00(7)
`-0.23 (4)
`
`Cation I
`-0.60(5)
`-0.65 (6)
`-0.86 (5)
`-1.91 (9)
`0.06 (4)
`Cation II
`-0.32 (4)
`-0.65 (4)
`0.28 (4)
`0.23 (6)
`-0.19 (4)
`Anion
`-0.12(3)
`o,
`-0.19 (3)
`-0.19 (3)
`2.14 (4)
`1.79
`1.85 (3)
`1.49(3)
`0.25 (4)
`2.21
`0.07 (3)
`0.28 (4)
`4.73 (7)
`1.33 (3)
`1.75 (4)
`o2
`-0.00(3)
`0.15 (3)
`0.25 (2)
`2.35 (4)
`1.87
`1.82 (3)
`1.57(3)
`03
`-0.82(4)
`o,
`-0.04 (3)
`2.19
`0.06(3)
`2.76 (4)
`3.08 (6)
`1.35 (3)
`-0.07 (1)
`0.09 (1)
`0.04 (1)
`1.38 (1)
`1.41
`1.41 (1)
`1.46(1)
`P
`" The number in parentheses that follows each A„ value is the estimated standard deviation in the last significant figure. The B,/s in A2
`b Atoms numbered to agree with Figures 1-5.
`Iso-
`are related to the dimensionless ß,, employed during refinement as B,¡ = 4ß,, ,«*«,*.
`tropic thermal parameter calculated from B = 4[K2 det(j3,·,·)]'/2
`
`2.40
`2.77
`2.75
`3.48
`2.07
`
`2.52
`2.49
`2.62
`3.39
`2.08
`
`1
`
`Table III. Refined Parameters for Hydrogen Atoms in Crystalline
`Bis(methylguanidinium) Monohydrogen Phosphate"
`
`Atom6
`type
`
`Hm
`Hn21
`Hn22
`Hn si
`Hn32
`Hen
`Hci»
`Hci3
`
`Hni
`Hn21
`Hn22
`Hn31
`Hn32
`Hen
`Hen
`Hci3
`
`103x
`
`99(1)
`101 (1)
`85(1)
`54 (2)
`61 (1)
`138 (2)
`157 (1)
`102 (2)
`
`200(1)
`309 (1)
`296 (1)
`211 (1)
`175(1)
`256 (2)
`309 (2)
`278 (2)
`
`Isotropic
`thermal
`parameter,
`fl, Á2
`
`0.7 (5)
`1.5(6)
`1.5(6)
`2.1 (7)
`0.2 (4)
`4.1 (9)
`2.1 (7)
`5.3(11)
`
`0.3 (4)
`0.9 (5)
`1.3(5)
`1.1 (5)
`1.0(5)
`7.6(16)
`5.8 (13)
`5.4 (14)
`
`103z
`
`-262 (4)
`-646 (4)
`-731 (5)
`-570 (5)
`-399 (4)
`-243 (7)
`-415(5)
`-393 (7)
`
`117(3)
`310(4)
`246 (5)
`112(4)
`54 (4)
`196(8)
`128 (7)
`322(8)
`
`103y
`Cation 1
`85(1)
`139 (1)
`83 (1)
`12(1)
`8(1)
`162(2)
`159(1)
`189 (2)
`Cation II
`7(1)
`66(1)
`127 (1)
`138 (1)
`92(1)
`-62(3)
`-25(2)
`-26(2)
`Anion
`-138 (11)
`0.0
`Ho,
`0.0
`5.8 (19)
`H0,
`0.0
`0.0
`8.8 (25)
`143 (12)
`“ Figures in parentheses are the estimated standard deviations of
`the last significant digit. Coordinates listed without standard de-
`6 Atoms numbered to agree with
`viations are symmetry required.
`Figures 1-5.
`
`ordered. The experimental data did not allow a choice;
`both the truly symmetric structure or one in which there
`are nearly symmetric but disordered bonds are
`con-
`sistent with the data. However, Speakman, in a recent
`review of short hydrogen bonds,21 has concluded that
`likely only when the
`symmetrical hydrogen bonds are
`O-H-O distance is less than 2.44 A, and on this basis
`we believe that the hydrogen bonds in this structure are
`unsymmetrical and disordered. Should these hydrogen
`the re-
`lower temperatures,
`atoms become ordered at
`(28) J. C. Speakman, Struct. Bonding (Berlin), 12, 141 (1972).
`
`Figure 1. A perspective view of one asymmetric unit. The atom
`numbering scheme is explained in the text.
`For clarity the actual
`thermal ellipsoids of the hydrogen atoms ae not used.
`
`Discussion
`Overall Structure. As emphasized in Figure 3,
`the
`structure can be thought of as centered around a pair
`of monohydrogen phosphate ions which are
`held
`together by two strong hydrogen bonds. The O- · O
`distances in these bonds are 2.544 and 2.503 A which
`that they are close to or perhaps within the range
`means
`where the hydrogen bonds might be symmetrical.
`Moreover, there is a crystallographic twofold axis which
`passes through the midpoints of
`the two hydrogen-
`bonded 0·· 
`this symmetry re-
`pairs. However,
`quirement could be satisfied either by having the hydro-
`gen atoms on the twofold axis (truly symmetrical hydro-
`gen bonds) or by having them off the axis but dis-
`
`Journal of the American Chemical Society / 96:14 / July 10, 1974
`
`Merck Exhibit 2208, Page 4
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`Table IV. Bond Lengths in Bis(methylguanidinium) Monohydrogen Phosphate"
`.,2
`Cation II
`l .448 (3)
`1.330 (2)]
`1.327 (2) i
`1.331 (2)J
`
`Type6
`Ci-Ni
`Q-N,
`Q-H
`Q-Ns
`Q-N
`Av
`N,-Hni
`N2-Hn21
`N2   22
`N3-Hn31
`N3—Hn32
`
`~
`
`,1
`
`Cation I
`1.442 (3)
`1.322 (3)
`1.324 (3)
`1.328 (2)
`
`1.325
`0.82(3)
`0.91 (3)
`0.87 (4)
`0.79 (4)
`0.84 (3)
`
`1.329
`0.86 (3))
`0.85 (3)|
`0.96(3)1
`0.82(3)|
`0.89 (3)J
`
`Av
`1.445
`
`1.327
`
`Type6
`Ci-Hc,
`Cl-Hci2
`Cl-Hci3
`
`P-O,
`
`0.86
`
`P-O,
`
`p-o3
`p-o4
`
`4475
`
`Av
`
`0.96
`
`1.562
`(1.569)=
`
`1.519
`(1.527)=
`
`1 27
`
`1
`
`1
`
`j
`
`1
`
`Cation I
`Cation II
`0.93 (6)1
`0.84(5)
`1.04 (5))
`0.93 (4)
`0.96(6)j
`1.07 (5)
`Anion bond length, A
`1.567 (1)
`(1.573)=
`1.556 (1)
`(1.565)=
`1.514(1)1
`(1.521)=
`1.524(1)1
`(1.533)=
`)
`1.276 (7)
`O1-H01
`1.263 (13)j
`O2-H02
`b Atoms numbered to agree with Fig-
`" The figure in parentheses following each individual distance is the estimated standard deviation.
`= Bond length corrected for libration of HP042~ group as a rigid body according to V. Schomaker and
`ures 1-5 and Tables I and III.
`K. N. Trueblood, Acta Crystallogr., Sect. B, 24,63 (1968).
`
`suiting crystals of this compound could prove to be
`ferroelectric.29
`Surrounding each central dimer of phosphate ions is
`a total of 12 methylguanidinium ions, six from each of
`the two crystallographically distinct cations.
`These
`are linked to the phosphate dimer by a total of 18, 20,
`or 22 hydrogen bonds, namely, ten from cation II and
`8, 10, or 12 from cation I. The hydrogen bonding pat-
`tern of a single guanidinium ion is illustrated in Figure
`4 for cation I and in Figure 5 for cation II. Considering
`a single phosphate dimer, cation I and II both form a
`single hydrogen bond to Os of each phosphate ion for a
`total of four. Cation II
`forms two pairs of hydrogen
`the phosphate dimer, giving al-
`bonds bridging across
`together four of this pattern and a subtotal of eight.
`Cation I and cation II each form a pair of hydrogen
`bonds to two oxygen atoms of one phosphate ion, giving
`eight H bonds of this type for a subtotal of 16. This
`type of guanidinium-phosphate interaction, which is
`best illustrated in Figures 1 and 2, has also been observed
`in the structure
`of methylguanidinium dihydrogen
`orthophosphatelbi 23 and in the structure of propyl-
`the most
`guanidium diethylphosphate.30 However,
`significant aspect of this type of paired hydrogen bond
`interaction between a guanidinium ion and a phosphate
`ion is that it provides an excellent model for the inter-
`action of arginines-35 and -87 of the Staphylococcal
`nuclease with the 5'-phosphate of its potent inhibitor,
`thymidine 3',5'-diphosphate, as illustrated diagramati-
`cally by 2.
`The final type of guanidinium-phosphate interaction
`observed in this structure is the   -- —02 of cation I
`(Figure 4) with an N-O distance of 2.90 A which is
`paired with another N-O interaction to a different oxy-
`i.e., N!2--H--04 with
`gen atom of the same phosphate,
`an N-O distance of 3.17 A. This distance is somewhat
`long to be considered a real hydrogen bond, and the
`final N-O interaction shown in Figure=4, that of Ni2~
`H--02 with an N-O distance of 3.27 A, is even more
`doubtful.
`The grand total of N to O hydrogen bond-
`ing interactions to a single dimer of phosphate ions is
`thus 18, 20, or 22, the latter two numbers including one
`or both of the pair of N-O interaction over
`3 A. This
`(29) W. C. Hamilton and J. A. Ibers, “Hydrogen Bonding in Solids,”
`W. A. Benjamin, New York, N. Y., 1968, pp 238-255.
`(30) S. Furberg and J. Solbakk, Acta Chem. Scand., 26, 3699 (1972).
`
`CATION I.
`
`Figure 2. Bond lengths and angles for the asymmetric unit as seen
`A complete set of values and standard deviations
`in Figure 1.
`are listed in Tables IV, V, and VI.
`
`final pattern of guanidinium-phosphate interaction,
`showing in this structure a third paired cyclic N-O sys-
`tem formed by one strong hydrogen bond and a second
`rather weak one, may well have some
`biochemical
`In the structure of propylguanidinium
`significance.
`diethylphosphate30 one paired interaction with two
`strong hydrogen bonds is observed, but there is also a
`second pair with the one strong and one weak pattern
`that we also observe in this structure.
`The Staphylo-
`coccal nuclease is, of course, a phosphodiesterase with
`interaction with
`Its initial
`a degree of base specificity.
`
`Cotton, et al.
`
`/ Bis(methylguanidinium) Monohydrate Orthophosphate
`
`Merck Exhibit 2208, Page 5
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`4476
`Table V. Bond Angles in Bis(methylguanidinium) Monohydrogen Phosphate"
`
`Type6
`Ni-Ci-Ns
`Ni-Ci-N,
`n2-c2-n3
`Q-Ni-C2
`Hm-Nj-Q
`HNi-N!-C2
`
`Hn21-N2-Hn22
`Hn21~N2~C2
`Hn22-N2-C2
`Hn31-N3-Hn32
`Hn31~Ns~C2
`HN32-N3-C2
`
`Cation I
`121.4(1)
`119.4(2)
`119.2(2)
`124.0(2)
`115 (2)
`119(2)
`
`123 (3)
`120 (2)
`117 (2)
`115 (3)
`121 (3)
`123 (2)
`
`Cation II
`120.7 (2)
`118.8 (2)
`120.5 (2)
`123.9 (2)
`120(2)
`114(2)
`
`122(3)
`120 (2)
`118 (2)
`119 (3)
`120 (2)
`119 (2)
`
`Av
`Type6
`Hci-CrN,
`121.1
`Hci2-Cl-Nl
`119.1
`Hc,3-CrNl
`119.9
`Hcn-Cl-Hci2
`124.0
`Hcn-Ci-Hcu
`118
`Hci2-Ci-Hcic
`117
`Angles in Anion, deg
`Oi-P-02
`123
`Oi-P-O,
`120
`Ol-P-O,
`118
`o2-p-o3
`117
`o2-p-o4
`121
`o3-p-o.
`121
`P-O1-H01
`P-02-Ho2
`" Figures in parentheses are the estimated standard deviations of the last significant digit.
`and Tables 1 and III.
`
`Cation I
`109 (3)
`112(2)
`114(3)
`109 (4)
`114 (4)
`98 (3)
`
`Cation II
`
`111 (3)
`114 (3)
`104 (3)
`112(4)
`98 (4)
`117 (4)
`
`Av
`
`110
`113
`109
`
`111
`106
`108
`
`107.32 (6)
`109.33 (7)
`109.47 (7)
`109.35 (8)
`109.19 (9)
`112.07 (7)
`109 (3)
`112(4)
`6 Atoms numbered to agree with Figures 1-5
`
`Table VI.
`
`Donor atom (D)
`
`Asymmetric unit of (A)
`
`x, y, z
`x, y, z
`
`x, y, z
`x, y, z
`
`-
`
`ft + z
`
`x, 7» + y,
`y4 —
`x, y, -1 + z
`x, y, -1 + z
`x, y, -1 + z
`
`-
`
`V2 + z
`Vi
`-y,
`x,
`y, Vi + z
`1/i + x, 7< -
`lU + z
`lU + y,
`x,
`V4 —
`
`10(2)
`7(2)
`
`14(2)
`5(2)
`
`15(2)
`21 (2)
`27 (2)
`19(3)
`
`15(2)
`19(2)
`12(2)
`
`N!-HN1
`N3—HN32
`Ni-Hni
`N3-Hn32
`
`N2-Hn21
`N2-Hn22
`N2-HN22
`N3-Hn3i
`
`N2-Hn21
`N2—Hn22
`N3-Hn3,
`
`04
`o1
`
`03
`04
`
`o3
`o4
`02
`o2
`
`Os
`04
`Ol
`
`2.812 (2)
`2.935 (2)
`
`2.791 (2)
`2.905 (2)
`
`2.837 (2)
`3.169 (3)
`3.268 (3)
`2.900 (3)
`
`2.908 (2)
`2.865 (2)
`2.990 (2)
`
`Interionic Hydrogen Bonds in Bis(methylguanidinium) Monohydrogen Phosphate
`Distance, Á6
`Angle, deg6
`Distance, A6
`Acceptor
`H · A
`H-D ·
` 
`A
`D
`atom (A)
`Intracomplex Bonds. Cation Ic
`2.01 (3)
`2.11 (3)
`Intracomplex Bonds, Cation II«
`1.96(3)
`2.02(3)
`Intercomplex Bonds, Cation F
`1.97 (3)
`2.38 (4)
`2.52 (3)
`2.17 (4)
`Intercomplex Bonds, Cation IF
`2.10 (3)
`1.99 (3)
`2.19 (3)
`Intercomplex Bonds, Anionc
`5(4)
`2.544 (3)
`1.276 (7)
`O,
`— x, —y, z
`O1-H01
`8(4)
`1.263 (13)
`2.503 (3)
`-y,
`O2-H02
`-x,
`O»
`z
`6 Figures in parentheses are the estimated standard deviations of the
`“ The hydrogen atom actually involved in the bond is also indicated.
`c The complex is considered to be the asymmetric unit of Tables I and III and of Figures 1 and 2. All donor atoms
`last significant digit.
`belong to this asymmetric unit.
`
`the substrate through the agency of its two active-site
`guanidinium groups might fall into the two strong, one
`weak, one strong pattern of H bonds, found in propyl-
`If, as we have sug-
`guanidinium diethylphosphate.
`gested,20 enzymatic hydrolysis proceeds by the nucleo-
`philic back-side attack of a water molecule or hydroxide
`ion coordinated to the calcium ion, then, in the process,
`the guanidinium-phosphate interaction could be trans-
`formed easily to the two strong, two strong pattern of
`interaction, as observed in this structure, with a con-
`comitant increase of the positional and chemical stabil-
`ity of the transition state and consequent reduction of
`the energy barrier to reaction.
`In the nuclease,
`the
`observed geometrical arrangement of
`the guanidyl
`groups around the 5'-phosphate of the inhibitor appears
`to be consistent with such a notion.19·20
`Including the hydrogen bonds,
`the bonding arra